re: Using Combining Double Breve and expressing characters perhaps as if struck out.

From: verdy_p (verdy_p@wanadoo.fr)
Date: Sat Jul 24 2010 - 11:14:41 CDT

  • Next message: Philippe Verdy: "Re: Using Combining Double Breve and expressing characters perhaps as if struck out."

    > De : vanisaac@boil.afraid.org
    > Guys, does nobody read the bloody Standard anymore!?
    >
    > You CAN currently add a diacritic on top of a double diacritic. The "other"
    > base character is called the Combining Grapheme Joiner (U+304F).

    Sorry, I had forgotten this one. Note that I was not sure about the character to use as the base for additional
    diacritics (so I indicated « U+xyzt »).

    And I did not ask for encoding a new character, as I was nearly certain that such a solution existed using a base
    character with combining class 0.

    Ok, ZWJ was a bad guess, but *does* CGJ enter in the definition of « default grapheme clusters », or at least in the
    definition of « extended grapheme clusters » ? I hope it does, and that software are ready to support it as
    documented.

    Very few softwares were updated to support the still « recent » version 5.0 Unicode specifically, there are tons
    that still know and implement only Unicode 4.0. When rendering texts containing some CGJ, they will try to map it
    into fonts (where it will most often not be found, because old renderers typically also use old fonts), so they will
    display a .notglyph rectangle before the diacritic displayed with a dotted circle (as if it was starting a «
    defective sequence ») instead of being smarter and trying to place the diacritic on top of the previously seen
    cluster, or at least on top of the last character of the sequence containing the double diacritic...

    I'm so used to see all the defects in softwares based on Unicode 3.2 or 4.0 that I often forget that thre may exist
    newer solutions.

    Note that even Windows 7 does not include this CGJ control in its IME for rich text input controls, where it
    currently allows selecting and entering ZWJ and ZWNJ, or the various selector controls for « national » digit
    shapes, or the non-recommended BiDi embedding controls (or the really deprecated RS and US ASCII controls that could
    be more easily typed directly on the standard keyboard layout usng the Ctrl key in terminal emulators that still
    need these controls, but that are absolutely not needed in texts...).

    So I'm not alone to have forgotten it, Microsoft also forgot it for the standard text input controls in Windows 7,
    and browsers also completely forgot to include such selector facility for input elements.

    Philippe.



    This archive was generated by hypermail 2.1.5 : Sat Jul 24 2010 - 11:19:37 CDT